Inserting common functionalities

imxml2tei.py

@@ -7,9 +7,93 @@
 
 """
 
+import argparse
 import sys
 import configparser
+from pathlib import Path
 from lxml import etree
+import utils.libeoaconvert as libeoaconvert
+
+BASE_DIR = Path( __file__ ).resolve().parent
+SCRIPT_PATH = Path( __file__ )
+SCRIPT_NAME = SCRIPT_PATH.stem
+
+#####################
+# Parsing arguments #
+#####################
+
+parser = argparse.ArgumentParser(
+        formatter_class=argparse.ArgumentDefaultsHelpFormatter
+)
+parser.add_argument(
+        "-c", "--config",
+        default = BASE_DIR / "config" / "eoaconvert.cfg",
+        help="Name of config file"
+)
+parser.add_argument(
+        "-l", "--log-dir",
+        default = Path("output/logs"),
+        # default = Path("logs", SCRIPT_NAME).with_suffix(".log"),
+        help="logfile"
+)
+parser.add_argument(
+        "--log-level",
+        default = "INFO",
+        help="log level: choose between DEBUG, INFO, WARNING, ERROR, CRITICAL"
+)
+parser.add_argument(
+        "-f", "--filename",
+        default = "IntermediateXMLFile.xml",
+        help="Name of intermediate XML file (without suffix!)."
+)
+parser.add_argument(
+        "-o", "--output-dir",
+        default = "./output/tei",
+        help="where to dump all output files"
+)
+parser.add_argument(
+        "-i", "--input-dir",
+        default = "./output/imxml",
+        help="location of intermediate XML file"
+)
+
+args = parser.parse_args()
+
+CONFIG_FILE = args.config
+
+print("The configfile is %s." % CONFIG_FILE)
+
+CONFIG = load_config(
+        CONFIG_FILE,
+        args.log_level,
+        (Path(args.log_dir) / SCRIPT_NAME) . with_suffix( ".log" ),
+        # args.log_file,
+)
+
+############################
+# Paths:
+############################
+INPUT_DIR = Path( args.input_dir )
+INPUT_PATH = Path( args.filename )
+OUTPUT_DIR = Path( args.output_dir )
+LOG_DIR = Path( args.log_dir )
+
+TEMP_DIR = OUTPUT_DIR / "tmp_files"
+DEBUG_DIR = OUTPUT_DIR / "debug"
+
+# where to output the xml file:
+XML_FILE = (OUTPUT_DIR / INPUT_PATH.name) .with_suffix( ".xml" )
+
+##################################
+# Setting up various directories #
+##################################
+
+if not os.path.exists(OUTPUT_DIR):
+    os.mkdir( OUTPUT_DIR )
+if not os.path.exists(TEMP_DIR):
+    os.mkdir( TEMP_DIR )
+if not os.path.exists( DEBUG_DIR ):
+    os.mkdir( DEBUG_DIR  )
 
 # citations need a little more work: especially citedRange
 # so do landscape figures, no way to distinguish them!
@@ -417,10 +501,9 @@ def main():
     back_part = etree.SubElement(tei_body, "back")
     tei_body.insert(1, tei_body_xml.getroot())
 
-    outfile = 'CONVERT/TEI.xml'
     output_string = etree.tostring(tei_root, xml_declaration=True, pretty_print=True, encoding="UTF-8", doctype= '<?xml-model href="eoa_tei.rnc" type="application/relax-ng-compact-syntax"?>\n<?xml-stylesheet type="text/css" href="tei.css" ?>')
 
-    with open(outfile, 'w') as output_file:
+    with open(XML_FILE, 'w') as output_file:
         output_file.write(output_string.decode("utf-8"))
 # def main ends here